Post on 21-Jun-2020
transcript
Differential Gene Expression in Danio rerio during
Optic Nerve Regeneration
THESIS
Presented to the Graduate Council of Texas State University-San Marcos
in Partial Fulfillment of the Requirements
for the Degree
Master of SCIENCE
by
Katherine E. Saul, B. S.
San Marcos, Texas, August 2008
Differential Gene Expression in Danio rerio during
Optic Nerve Regeneration
Committee Members Approved: Dana M. García, Thesis co-Chair Joseph R. Koke, Thesis co-chair
Nihal Dharmasiri Timothy Raabe
Approved:
____________________________________ J. Michael Willoughby Dean of the Graduate College
iii
ACKNOWLEDGEMENTS
I am eternally grateful for all the support of my friends and family. The guidance
of my professors, Dr. Koke and Dr. García, has been inspirational, and I look forward to
working together in the future. I am very appreciative to Jeff Landgraf at Michigan State
University for always being so willing to assist me in understanding the microarray data.
I also express my gratitude to Angela Archer at Eppendorf for all of the assistance with
quantitative PCR. This thesis would not have been possible without the surgery team,
late night pipetting, last minute image acquisition, and concept mapping over coffee with
my lab mates. To Amanda Mosier, Elizabeth Capalbo, Mayuri Patel, and John
Miller…THANKS!
The inspiration for this project was fueled by a spinal cord injury that left my
brother, Devon, partially paralyzed. Devon, thanks for not listening to the doctors who
said you’d never walk again. You’ve beaten the odds and given hope to all of us.
iv
TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS......................................................................................... iii
LIST OF TABLES....................................................................................................... v
LIST OF FIGURES .................................................................................................... vii
ABSTRACT................................................................................................................ viii
CHAPTER
I. INTRODUCTION.........................................................................................1
II. MATERIALS AND METHODS.................................................................4
III. RESULTS ...................................................................................................12
IV. DISCUSSION ............................................................................................25
APPENDIX TABLES ..................................................................................................34
REFERENCE LIST .....................................................................................................44
v
LIST OF TABLES
Table Page
1. Gene Selection and Primer Design ................................................................................10
2. Genes Differentially Expressed by 1.5-fold or more .....................................................13
3. Microarray Results: Genes Showing Increased Expression at 3 Hours.........................34
4. Microarray Results: Genes Showing Increased Expression at 24 Hours.......................35
5. Microarray Results: Genes Showing Increased Expression at 168 Hours.....................36
6. Microarray Results: Genes Showing Decreased Expression at 3 Hours .......................37
7. Microarray Results: Genes Showing Decreased Expression at 24 Hours .....................38
8. Microarray Results: Genes Showing Decreased Expression at 168 Hours ...................39
9. Gene Ontology: Cell Proliferation ................................................................................40
10. Gene Ontology: Axon Extension and Guidance.........................................................41
11. Gene Ontology: Embryonic Development..................................................................41
12. Gene Ontology: Immune System Response ...............................................................42
13. Gene Ontology: Neuron Differentiation ......................................................................42
14. Gene Ontology: Phototransduction..............................................................................43
15. Lunatic Fringe qRT-PCR Results ................................................................................17
16. Hoxa11b qRT-PCR Results .........................................................................................18
17. ATF3 qRT-PCR Results ..............................................................................................20
18. YY1 qRT-PCR Results ................................................................................................21
19. Nog2 qRT-PCR Results...............................................................................................22
vi
LIST OF TABLES (Continued)
Table Page
20. Tubb5 qRT-PCR Results .............................................................................................23
21. KLF7a qRT-PCR Results ............................................................................................24
vii
LIST OF FIGURES
Figure Page
1A. Optic Nerve Injury Methods .......................................................................................6
1B. Optic Nerve Injury Methods .......................................................................................6
2. Temporal Analysis of Gene Ontology During Optic Nerve Regeneration..................13
3. Temporal Expression of Genes Involved in Phototransduction ..................................16
viii
ABSTRACT
Spinal cord injuries and neurodegenerative diseases in mammals result in a loss of
function due to the failure of neurons in the central nervous system (CNS) to survive and
regenerate their axons. Unlike mammals, fish and amphibians possess the ability to
regenerate their CNS following damage. To gain a better understanding of the factors
necessary for successful CNS regeneration, I conducted a temporal analysis of the
changes in gene expression in the retina caused by optic nerve injury to identify genes
specifically involved in regeneration. Dual color oligonucleotide microarrays were used
to compare total RNA harvested from retinas of sham-operated and optic nerve-injured
fish at 3, 24 and 168 hours following surgery. Statistical analyses identified 722 genes
differentially expressed by at least 1.5-fold at one or more time points, and 142 genes
with at least a 2.0-fold difference. Based on microarray fold differences and gene
ontology analysis, six genes were selected for further analysis using qRT-PCR. The
results of qRT-PCR identified noggin 2, activating transcription factor 3, and beta-tubulin
5 as genes that showed significantly increased expression in the injured fish as compared
to sham; therefore, these genes’ products may play an important role in optic nerve
regeneration in zebrafish. These results support the hypothesis that an analysis of gene
expression between optic nerve injured and sham-operated fish will reveal genes
specifically involved in regeneration.
1
CHAPTER I
INTRODUCTION
Spinal cord injuries and neurodegenerative diseases in mammals result in a loss of
function due to the failure of neurons in the central nervous system to survive and
regenerate their axons. At some point during the development of mammals, the central
nervous system (CNS) neurons lose their ability to regenerate axons and reestablish
functional connections after axotomy (Cho et al., 2005). The retina and optic nerve are
developmentally and functionally part of the brain (and CNS), and because of its
accessibility, optic nerve injury has become a standard model system for studies of nerve
regeneration in the CNS.
The axons that make up the optic nerve originate from ganglion cells in the retina
and project primarily to the optic tectum of the brain. In mammals, damage to the optic
nerve results in wallerian degeneration of axons distal to the injury site (towards the
brain) and apoptosis of cell bodies in the retina. Growth of any new neurites that may
sprout from the nerve stump is attenuated by both physical barriers, specifically the
formation of the glial scar (Goldberg and Barres, 2000; Ries et al., 2007) and molecular
barriers such as myelin inhibitory protein, semaphorin 3A, and chondroitin sulfate
proteoglycans (Cao et al., 2008).
2
In contrast, it has been known for many years that fish and amphibians possess
the ability to spontaneously regenerate axons in the central nervous system following
axotomy. Fish possess the same cellular and molecular players that prevent regeneration
in mammals, but following nerve injury, the outcome is much different. Regenerating
neurites sprout from the nerve stump and functionally re-enervate the brain (Attardi and
Sperry, 1963; Gaze and Keating, 1972; Gaze et al., 1972; Veldman et al., 2007).
Why does this perplexing difference between fish and mammals exist? To
approach this question and gain a better understanding of the factors necessary for
successful CNS regeneration, an examination of the changes in gene expression during
optic nerve regeneration in zebrafish was conducted. The working hypothesis for this
experiment was that observing differences in gene expression between optic nerve
injured and sham-operated fish will reveal genes involved in regeneration. A better
understanding of the genetic mechanisms behind this remarkable capability will provide
understanding of the requirements for functional recovery after nerve trauma.
Zebrafish (Danio rerio) are an excellent model system for studying nerve
regeneration for several reasons: they are inexpensive, easily maintained, and their
genome is fully sequenced. To observe changes in gene expression during optic nerve
regeneration, RNA extracted from injured and sham-operated zebrafish eyes was
compared using oligonucleotide microarray analysis. Microarrays allow the
simultaneous examination of the expression of a large number of genes - in this case
approximately 9,000 (14,067 unique ~50-mer probes representing 8,839 genes) at each
time point. The results of the microarray assays were statistically analyzed using
3
bioinformatics software, and the change in expression of genes identified as suspects was
further analyzed using quantitative PCR.
4
CHAPTER II
MATERIALS AND METHODS
Fish maintenance
Wild-type zebrafish were obtained from a local pet store (Animal Wonders, San
Marcos, TX). Fish were conditioned on a 12 hour light/dark cycle for a minimum of 14
days before use. All protocols were approved by the Texas State IACUC (approval #
0703_0122_07).
Experimental Design
The experimental design of this analysis compared three interventions designated:
the injured, sham-operated, and control. The surgical procedures for each are described
in the section ”Optic Nerve Injury.” For this experiment, I used a dual-color microarray
to compare between the sham-operated and optic nerve-injured fish. Previous studies of
this nature have examined injured retina as compared to control retina. With this
approach genes involved in general tissue restoration and immune response may also
show significant temporal change and thus confound the analysis of gene expression
changes important to nerve regeneration. By comparing injured to sham-operated fish, I
attempt to dissect out the “noise” of non-neuronal tissue repair and inflammatory
response, while emphasizing gene responses specific to neural injury and repair.
5
In general, a fold increase in an experimental sample as compared to a control
would indicate an up regulation in gene expression. The experimental design employed
in this study compares the experimental to the sham instead of a control; therefore, I am
unable to determine whether the gene expression ratios greater than 1 are due to an
increase in gene expression in the injured fish, decreases in gene expression in the sham-
operated fish, or a differential increase or decrease in both. Thus, genes of interest
selected from the microarray analysis were further analyzed by qRT-PCR where samples
from experimental and sham retinas were compared to control retinas.
Total RNA was isolated from sham-operated and optic nerve-injured retinas at 3
time points. I selected 3 hours, 24 hours and 168 hours to compare early changes in gene
expression (3 and 24 hours) to subsequent changes in expression as regenerating axons
are first observed synapsing in the brain (168 hours) (Bernhardt et al., 1996).
Optic Nerve Injury
Optic nerve injury was performed as described below using a method modified from
Q. Liu and R. L. Londraville 2003 (Liu and Londraville, 2003). The zebrafish were
anesthetized in 0.2% Finquel® tricaine methanesulfonate (MS-222, Argent Chemical
Laboratories, Redmond, Washington) dissolved in tank water. The zebrafish were
wrapped in a wet paper towel exposing only the head, and placed on a stereomicroscope
for dissection. Surgical tools were sterilized with 70% ethanol. By separating the dorsal
connective, cutting the lateral rectus muscle, and then angling the eye rostrally we are
able to expose the optic nerve. Taking care not to damage the ophthalmic artery, the
optic nerve was partially severed (~90%) using 3mm microscissors (EM Sciences,
Hatfield, PA) (Figures 1A and 1B). The eye was placed back into the socket and the fish
6
revived by being placed in aerated aquarium water. Sham operations were identical
except the optic nerve was not severed. Control fish were un-operated.
Figure 1A. Optic Nerve Injury Methods. First, the connective tissue surrounding the dorsal aspect of the eye was separated using a scalpel (A) and the lateralis muscle of the eye was cut using microscissors (B). The eye was angled slightly to expose the optic nerve which was then severed approximately 90% using microscissors (C) (see also Figure 1B below). This illustration adapted from Liu and Londraville (2003).
Figure 1B. Optic Nerve Injury Methods. The extent of optic nerve injury (arrow) is shown in vivo in this image. The optic nerve travels from the retinal ganglion cells of the eye to the brain. During the optic nerve injury procedure, care was taken not to damage the ophthalmic artery. Image was captured using a Nikon SNZ 1500 dissecting scope equipped with a Nikon DXM1200C digital camera.
7
RNA Extraction
All fish were sacrificed at midday to avoid any gene expression differences
associated with diurnal rhythm. Following euthanasia by overdose in MS-222, whole
eyes were removed from the fish 3 hours, 24 hours, and 7 days after optic nerve injury or
sham operations and immediately placed in RNA later (Ambion; Austin, TX). To
achieve 10 µg of total RNA required for the microarray, the retinas from 10-15
identically treated fish were pooled. The sclera and the lens were removed, and the
remaining eye tissues (retina, RPE, and choroid) were placed in 1 ml of TRI-Reagent
(Ambion; Austin, TX). Samples at each time point were collected in triplicate. The tissue
was homogenized by trituration with a 27 gauge needle and syringe, and total RNA was
isolated by organic extraction and isopropanol precipitation. The RNA isolated from
RPE is often contaminated with pigment (Malik et al., 2003, Invest Ophthalmol Vis Sci,
44, 2730-5) and this was confirmed in this study. RNA clean-up was performed using
RNeasy spin columns (QIAGEN, Valencia, CA), resulting in a pigment free product.
RNA quality and integrity was assessed using a Nanodrop spectrophotometer
(Thermofisher Scientific, Waltham, MA) and glyoxal gel electrophoresis with ethidium
bromide staining to detect the 18S and 28S rRNA bands (Sambrook and Russell, 2001).
Samples comprising intact RNA as indicated by a lack of smearing on gels were sent to
Michigan State University’s Core Genomics Facility for an additional quality check using
the Agilent BioAnalyzer and subsequent microarray analysis.
Microarray Analysis
Microarray analysis was performed by Dr. Jeff Landgraf at Michigan State
University as follows. Labeling of the RNA for the oligonucleotide array was performed
8
using the Amino-Allyl MessageAMP II aRNA Amplification Kit (Ambion; Austin, TX).
During this procedure, total RNA is reverse transcribed using an oligo(dT) primer
bearing a T7 promoter using ArrayScriptTMas reverse transcriptase. The resulting cDNA
undergoes second strand synthesis in the presence of RNase H and subsequent
purification to serve as the template for in vitro transcription (IVT). During IVT,
modified nucleotides, 5-(3-aminoallyl)-UTP (aaUTP), are incorporated into the antisense
RNA (aRNA) during amplification. The aaUTP contain a reactive amino group on the
C5 position that can be coupled to N-hydroxysuccinimidyl ester-derivatized dyes (Cy3
and Cy5).
First strand cDNA synthesis was carried out on 1000 ng (1 µg) of total RNA at
42ºC for 2 hours. After second strand synthesis at 16°C for 2 hours, cDNA was purified
through a filter cartridge. IVT was carried out for 12 hours at 37ºC, and the resulting
aRNA was purified through a filter cartridge. The aRNA samples were coupled to either
Cy3 or Cy5 dye. The dyes were swapped between sham and experimental sample
replicates to control for any dye bias. This means that in one replicate sample, the sham
was labeled with Cy3 while the injured was labeled with Cy5, and in two replicates the
labels were switched. Each aRNA sample was fragmented using Ambion’s RNA
Fragmentation Reagents, added to the hybridization solution (Ocimum Biosolutions,
Indianapolis, IN), heated to 95ºC for 3 min, cooled on ice for 3 minutes, and spun briefly.
Hybridization was performed under a 40mm x 22 mm LifterSlip (Erie Scientific) by the
addition of 240 µl of labeled solution to Zebrafish 14K OciChipTM(Ocimum
Biosolutions). Slides were scanned using an Affymetrix 428 ArrayScanner and analyzed
with the GenePix Pro 3.0 software (Axon Instruments, Sunnydale, CA).
9
Array normalization and statistical analysis were performed using the “limma:
Linear Models for Microarray Data” library module (version 2.2.0) of the R statistical
package (version 2.2.0). Signal intensities close to the background are considered
unreliable data (Korenberg, 2007); therefore, all signal intensities of less than 1000 in all
times points were removed from the analysis. A one-way analysis of variance (ANOVA)
was performed on the ratios of the triplicate time points to determine whether statistically
significant differences existed among gene expression between sham and experimental
(statistically different from a ratio of 1). Slide intensity data were normalized using the
global LOWESS (locally weighted scatter plot smoothing) method with the least squares
method used for the linear model fit. The purpose of normalization is to remove
technical variation while still retaining biological signal (Gentleman et al., 2004).
Gene Ontology Analysis
Gene ontology provides a computational approach to answering questions such
as, “what is known about the biological function of these genes?” Gene ontology is a
controlled vocabulary that is used to describe knowledge and implications of the
biological process, molecular function, and cellular localization of gene products
(Korenberg, 2007 and Ashburner et al., 2000). The biological process is particularly
useful because up- or down-regulation of a set of genes within the same process provides
evidence that a specific cellular event has been activated. All genes that displayed
greater than 1.5-fold change in at least one time point were individually analyzed for
known functional characteristics and gene ontology, using the bioinformatics software
GeneSifter ® (VizXLabs, Seattle, WA), the web-based search engine GeneTools
(Beisvag et al., 2006), and ontology information provided by the chip manufacturer. In
10
the case of GeneSifter and GeneTools, ontology analysis was accomplished based on
gene accession number.
Quantitative PCR
Corroboration of the results of the microarray analysis was sought by quantitative
reverse transcriptase polymerase chain reaction (qRT-PCR). Based on microarray fold-
differences and gene ontology, seven genes were selected for validation (Table 1).
Primers for these genes were designed observing the following criteria: primer length
18-25 bases long, 50-60% GC content, and with melting temperatures within 5oC of each
other. The gene product size for qRT-PCR should lie between 80 and 250 bases long
(Bustin, 2004), and this was confirmed using UCSC’s In-Silico PCR
(http://genome.ucsc.edu/). Secondary structures and self-complementarity were assessed
using an online oligo-nucleotide calculator (http://www.basic.northwestern.edu) (Table
1).
Table 1. Gene Selection and Primer Design. Genes selected for validation using qRT-PCR and their respective forward and reverse primers. The annealing temperatures used in the PCR program cycle for each of the primer pairs are given (Tanneal).
Gene Forward Primer 5’3’ Reverse Primer 5’3’ Tanneal
Noggin 2 CGCTTCTGAAGTTCCGATTG CTGAGCAATGAGGCTCCAGC 56.2OC
Lunatic Fringe GCGTCTCATAGCAATGGCG GGCATAGTGATGTCCAACTG 53.7OC
Hox A-11 CTCGGTTCTCTACCACTCC TGTCCACCGGATGCTCAGTC 55.5OC
ATF3 CCTTGTCATCTCCACGTCCAC CAGACCTTCCTGCTCACAGC 53.7OC
TF YY1 AGACGACGACGAGCACCA CTTGCCAGACACGGTCAC 55.3OC
ß-Tubulin AAACCGCCGTCTGCGATATTCC ACTACCACCTCCCCAAAACACC 59.0OC
KLF7A CATTACGTCTCCTCTGTTGG AAAGATTGGGATTGCTGGCTTG 55.3OC
GAPDH CAAGGGGTCACATCTACTC TGGGTGCTGGTATTCTCTC 53.0OC
11
Quantitative RT-PCR was performed using the Express SYBR® GreenER™
One-Step qRT-PCR Universal Kit (Invitrogen, Carlsbad, CA) on 10 ng of total RNA with
a total reaction volume of 20 µl. The following cycling program was executed using an
Eppendorf Realplex2 Mastercycler (Hamburg, Germany): 5 min at 50oC for cDNA
synthesis, 95oC for 2 min, 40 cycles [95oC for 15 sec, gene specific annealing
temperatures (Table 1) for 15 sec, 20 sec for extension at 60oC], followed by a melt curve
analysis. A gradient analysis was performed on each primer pair to determine the
optimum annealing temperatures. The gradient temperature ranged from 0.5ºC below the
lower recommended Tm to 60ºC. The optimal temperature was determined by
examination of the amplification plots and selection of the temperature that with the
lowest CT value, indicating the most efficient reaction.
Analysis of qRT-PCR results was completed using the 2-ΔΔ CT method (Livak and
Schmittgen, 2001) where:
ΔΔ CT = (CT, Target – CT, HKG)Injury - (CT, Target – CT, HKG)Sham
or
ΔΔ CT = (CT, Target – CT, HKG)Injury or Sham - (CT, Target – CT, HKG)Control
By using the 2-ΔΔ CT method, the data are represented as the fold change in gene expression
normalized to an endogenous reference gene (HKG, or house-keeping gene) and relative
to a control (or sham in this instance). The reference gene chosen was glyceraldehyde-3-
phosphate dehydrogenase (GAPDH).
12
CHAPTER III
RESULTS
Microarray Analysis
Total RNA was isolated from sham-operated and optic nerve injured retinas at 3
time points. I chose 3 hours, 24 hours and 168 hours post surgery to compare early
changes in gene expression (3 and 24 hours) to subsequent changes in expression as
regenerating axons are first observed forming terminal arborizations in the optic tectum
(168 hours) (Bernhardt et al., 1996). The Zebrafish 14K OciChipTM Oligo-nucleotide
Array (Ocimum Biosolutions) comprises 14,067 unique ~50-mer probes representing
8,839 genes. Statistical analyses identified 722 genes differentially expressed by at least
1.5-fold in one or more time points, and 142 genes with at least a 2.0-fold difference.
Table 2 illustrates the changes in differentially expressed genes (1.5-fold or more) at each
time point. The 20 most differentially expressed genes at each time point are displayed in
Tables 3 through 8 (Appendix).
13
Table 2. Genes Differentially Expressed by 1.5-fold or more. Ratios were expressed as Injury/Sham so that genes that displayed more intense signal intensity in the injury model vs. the sham model would result in a ratio greater than 1. “Genes Up” refers to genes that showed greater expression in the injured vs. the sham-operated fish (ratio > 1.5) “Genes Down” refers to genes that showed greater expression in the sham vs. the injured fish (ratio < 0.67).
Time Point Genes Up Genes Down
3 Hrs 50 17
24 Hrs 112 86
168 Hrs 217 191
All genes that displayed greater than 1.5-fold change in at least one time point
were individually analyzed for known functional characteristics and gene ontology using
bioinformatics software (Figure 2 and Tables 9 through 14).
Figure 2. Temporal analysis of gene ontology during optic nerve regeneration. Total number of genes differentially expressed at least 1.5-fold within each ontological category. The smallest category represented was from the immune system response category that indicated the largest number of differentially expressed genes occurred at 24 hours.
14
At all time points, those genes that were differentially expressed were primarily in
the ontology category of cellular process, metabolic process, and gene expression.
Cellular process is a broad category including actions such as cell adhesion,
communication, homeostasis, and proliferation. Metabolic process includes genes
involved in biosynthetic and catabolic processes, generation of precursor metabolites and
energy, and primary and secondary metabolic processes. The category of gene
expression includes genes that are involved in nucleic acid binding, or regulate
transcription or translation.
Overall, there is an increase in expression of genes associated with cell growth
and proliferation (Table 9) as well as genes associated with axon extension (tubulins) and
guidance (neural adhesion molecule L1.2 and ephrin a4b) (Table 10) in the injured fish as
compared to the sham-operated fish. Examples of genes up-regulated in the cell
proliferation and differentiation categories included ATF3, jun B proto-oncogene (a
member of the AP-1 transcription factor family), and fibroblast growth factor 24. There
was an increase in expression in the injured fish of genes associated with central nervous
system development (Table 11), specifically YY1, frizzled 2 (fzd2), noggin 2 (nog2),
tumor protein p63 (tp63), and lunatic fringe (lfng). Of the nine main categories of gene
ontology listed in figure 2, the immune system process category had the least number of
genes represented. The largest representation in the immune response category occurred
at 24 hours with 2 genes showing increased expression in the injured fish as compared to
the sham-operated fish.
A temporal analysis of the gene ontology categories indicates differences between
early and late changes in gene expression. Differential gene expression that occurred
15
during the early time points (3 hours and 24 hours) consisted of genes involved in cell
differentiation and proliferation (hoxa11b, YY1, and p63), developmental process (fzd2,
YY1, insulin-like growth factor binding protein 2, Dicer1, p63, nog2, lfng), and also
cytoskeletal genes (annexin A2a and tubulins). In addition, there was an increase in the
number of genes involved in basic physiological processes such as protein metabolism,
gene expression, and localization. Differential gene expression that occurred during the
later time point (168 hours) consisted of 51 genes identified as representatives of protein
metabolism, 32 within translation, 18 within localization, and 10 within development.
There 57 genes differentially expressed that were involved in cell growth and
proliferation (inhibitor of growth family, member 3, tubulins, and YY1) and 7 genes
implicated in differentiation (tumor protein p63-like, zgc:103619, baculoviral IAP repeat-
containing 5a, neural adhesion molecule L1.2, transcription factor AP-2 alpha, fibroblast
growth factor 24, and hematopoietically expressed homeobox). In addition, sixteen genes
associated with phototransduction were differentially regulated between the injured and
sham-operated models (Table 14). All but one, retinaldehyde binding protein (rlbp1),
were down regulated at 168 hours (Figure 3); however, rlbp1 expression was essentially
unchanged between the sham-operated and injured models with an expression ratio of
1.19.
16
Figure 3. Temporal expression of genes involved in phototransduction. Log2 of the ratio (injury/sham) shows a decrease in expression by 168 hours.
qRT-PCR Results
Quantitative RT-PCR was conducted on triplicate samples to validate the
microarray data on 6 genes represented on the microarray and one gene (KLF7a) that was
determined to be essential to nerve regeneration by Veldman et al. (2007). For the
purposes of the qRT-PCR discussion comparisons between injured and sham-operated
fish will be designated (I vs. S), comparisons between sham-operated and control fish (S
vs. C), and comparisons between injured and control fish (I vs. C).
Lunatic fringe was chosen for further investigation based on its 2.31-fold
difference at 24 hours (I vs. S) on the microarray, and its known role as a regulator in the
notch-signaling pathway during the segmentation phase of development (Lai, 2004). At
3 hours, the microarray results were insignificant, however, qRT-PCR results show a 1.6-
-fold change (I vs. S) (Table 15). When compared to the control, there is a marked
‐2
‐1.5
‐1
‐0.5
0
0.5
1
1.5
3Hours 24Hours 168Hours
Log2Ratio(Injury/Sham
)
TemporalExpressionofGenesInvolvedinPhototransduction
17
increase in expression in both the sham-operated (23.1-fold) and the injured (37.9-fold)
fish (Table 15). At 24 hours, the microarray indicated a 2.31-fold difference (I vs. S).
The results of qRT-PCR did not agree with the microarray; it revealed a 1.3-fold
difference (I vs. S). Both sham-operated and injured fish show significant (>1.5) increase
when compared to control, 1.5-fold and 2.0-fold respectively. At 168 hours, microarray
data was insignificant as was the qRT-PCR results (1.4-fold decrease I vs. S). Both
sham-operated and injured fish were significantly increased, 2.4-fold and 1.7-fold
respectively, when compared to control; however, the sham-operated had a larger
increase than the injured, giving the overall I vs. S ratio a value less than one.
Table 15. Lunatic Fringe qRT-PCR Results. The results of qRT-PCR do not reflect the same 2.31-fold change at 24 hours as the microarray. The numbers represented for qRT-PCR reflect the fold change calculated with the ∆∆C T method employing GAPDH as the housekeeping gene. The standard error is also reported (± standard error). Comparisons between injured and sham-operated fish are designated (I vs. S), comparisons between sham-operated and control fish (S vs. C), and comparisons between injured and control fish (I vs. C). Ratios with values <1 were converted to fold change by inversion and indicated by a negative sign preceding the number.
Time Point Microarray
(I vs. S) qRT-PCR
(I vs. S) qRT-PCR (S vs. C)
qRT-PCR (I vs. C)
3 hours -- 1.6 ± 0.11 23.1 ± 0.82 37.9 ± 3.41
24 hours 2.31 1.3 ± 0.12 1.5 ± 0.02 2.0 ± 0.22
168 hours -- -1.4 ± 0.05 2.4 ± 0.15 1.7 ± 0.04
Hoxa11b was chosen for further investigation based on its 1.85-fold differential
expression (I vs. S) at 3 hours on the microarray. Additionally, I was interested in this
gene because it is within a family of homeobox genes that code for transcription factors
involved in axial patterning and development. Hoxa11b has been shown to play a role in
cell differentiation and pattern formation during development and limb and tail
18
regeneration of newts (Beauchemin et al., 1994) and in the regenerating fin of adult
zebrafish (Geraudie and Borday Birraux, 2003). Perhaps this gene is also involved in
optic nerve regeneration. The qRT-PCR results indicate that at 3 hours there is a non-
significant fold change of 1.18 (I vs. S) (Table 16). However, when we look at how the
expression compared to the control, it is apparent that both the sham-operated and injured
fish showed a greater than 2-fold increase in expression. This suggests that this gene may
be responding to the general stress and wound and not to the neuronal repair. At 24 hours
there is an observed change of 1.94-fold (I vs. S), which is due to the increased
expression of hoxa11b in the injured model (1.7-fold I vs. C). At 7 days, we see no
significant differential expression (I vs. S) on the microarray or in qRT-PCR. There is a
30% decrease in expression (I vs. S), however, both sham-operated and injured fish
showed increased expression of hoxa11b when compared to the control.
Table 16. Hoxa11b qRT-PCR results. The results of qRT-PCR do not reflect the same 1.85-fold change at 3 hours as the microarray. The numbers represented for qRT-PCR reflect the fold change calculated with the ∆∆CT method employing GAPDH as the housekeeping gene. The standard error is also reported (± standard error). Comparisons between injured and sham-operated fish are designated (I vs. S), comparisons between sham-operated and control fish (S vs. C), and comparisons between injured and control fish (I vs. C). Ratios with values <1 were converted to fold change by inversion and indicated by a negative sign preceding the number.
Time Point
Microarray (I vs. S)
qRT-PCR (I vs. S)
qRT-PCR (S vs. C)
qRT-PCR (I vs. C)
3 hours 1.85 1.2 ± 0.07 2.2 ± 0.02 2.7 ± 0.18
24 hours -- 1.9 ± 0.14 -1.1 ± 0.04 1.7 ± 0.17
168 hours -- -1.3 ± 0.10 1.7 ± 0.13 1.3 ± 0.08
19
ATF3 was chosen for further investigation because it was among the most
differentially expressed genes on the microarray; it indicated fold increases at 24 hours
(3.6 I vs. S) and 168 hours (2.13 I vs. S). It also was selected because its gene ontology
indicated that the gene products play a role in cell differentiation and proliferation.
Previous studies have shown that ATF3 showed an increase in expression in axotomized
retinal ganglion cells and in axotomized peripheral nerves of rats (Hunt et al., 2004). At
3 hours the microarray showed non-significant results, however, qRT-PCR showed a 2.5-
fold change (I vs. S) (Table 17). This is primarily due to an increased expression of ATF3
in the injured model, however the 2.5-fold change (I vs. S) is slightly exaggerated by a
decrease in expression in the sham-operated fish as indicated by the 20% decrease when
sham is compared to control (-1.2-fold S vs. C). At 24 hours the microarray showed a
differential expression of 3.60-fold (I vs. S). The qRT-PCR results show a 2.1-fold (I vs.
S) increase. When compared to control, it is apparent that this difference in expression is
due to an increase in ATF3 expression in the injured fish (1.7-fold I vs. C). Similar to the
3 hours sample, the 2.1-fold change (I vs. S) is slightly exaggerated by a decrease in
expression in the sham-operated fish as indicated by the 20% decrease when sham is
compared to control (-1.2-fold S vs. C). At 168 hours, the microarray revealed a 2.13-
fold change. qRT-PCR results show a different story at 168 hours: the results show
essentially no differential expression between the injured, sham, or control model.
20
Table 17. ATF3 qRT-PCR results. The results of qRT-PCR reflect a similar fold change at 24 hours, but not at 168 hours when compared to the microarray. The numbers represented for qRT-PCR reflect the fold change calculated with the ∆∆CT method employing GAPDH as the housekeeping gene. The standard error is also reported (± standard error). Comparisons between injured and sham-operated fish are designated (I vs. S), comparisons between sham-operated and control fish (S vs. C), and comparisons between injured and control fish (I vs. C). Ratios with values <1 were converted to fold change by inversion and indicated by a negative sign preceding the number.
Time Point Microarray
(I vs. S) qRT-PCR
(I vs. S) qRT-PCR (S vs. C)
qRT-PCR (I vs. C)
3 hours -- 2.5 ± 0.16 -1.2 ± 0.04 2.1 ± 0.04
24 hours 3.60 2.1 ± 0.3 -1.2 ± 0.04 1.7 ± 0.12
168 hours 2.13 1.3 ± 0.10 1.0 ± 0.08 1.2 ± 0.01
YY1 was chosen for further investigation based on the microarray data indicating
a 1.85-fold (I vs. S) increase at 24 hours and a 2.62-fold (I vs. S) increase at 168 hours.
Additionally, gene ontology indicated its involvement in cell proliferation, gene
expression and differentiation. YY1 codes for a ubiquitous and multifunctional zinc
finger transcription factor protein that can activate or repress gene expression depending
on its binding partners (e.g., histone deacetylase 1). It is in the GL1-Krüeppel gene
family, and it has been shown that YY1 regulates the expression of diverse genes (e.g.,
p53, c-myc, IFN-β, CREB) that are important for cellular activity (Kurisaki et al., 2003).
At 3 hours the microarray data were not significantly differentially expressed (I vs. S) and
this was confirmed with the qRT-PCR (-1.1 I vs. S) (Table 18). However, when
compared to the control model, both injured and sham show a marked increase in
expression (17.8 and 19.5-fold respectively). At 24 hours the microarray data revealed a
21
1.85-fold difference (I vs. S) that was not supported by the qRT-PCR results as indicated
by -1.4-fold change (I vs. S). However, this change is not due to a decrease in
expression in the injured fish. Both the sham-operated and injured fish showed an
increase in expression of YY1 when compared to the control; the (S vs. C) fold change of
2.0 was greater than the (I vs. C) fold change of 1.5, thus giving the (I vs. S) ratio a value
less than 1. At 168 hours the microarray data showed a 2.62-fold difference (I vs. S), and
similar to 24 hours, was not supported by the qRT-PCR results as indicated by -1.2-fold
change (I vs. S). At all time points, YY1 was expressed at a higher level in the sham-
operated fish than in the injured.
Table 18. YY1 qRT-PCR results. The results of qRT-PCR do not reflect the 1.85-fold change at 24 hours (I vs. S) and the 2.62-fold change (I vs. S) at 168 hours reveal on the microarray. The numbers represented for qRT-PCR reflect the fold change calculated with the ∆∆CT method employing GAPDH as the housekeeping gene. The standard error is also reported (± standard error). Comparisons between injured and sham-operated fish are designated (I vs. S), comparisons between sham-operated and control fish (S vs. C), and comparisons between injured and control fish (I vs. C). Ratios with values <1 were converted to fold change by inversion and indicated by a negative sign preceding the number.
Time Point Microarray
(I vs. S) qRT-PCR
(I vs. S) qRT-PCR (S vs. C)
qRT-PCR (I vs. C)
3 hours -- -1.1 ± 0.04 19.5 ± 1.13 17.8 ± 0.87
24 hours 1.85 -1.4 ± 0.08 2.0 ± 0.09 1.5 ± 0.23
168 hours 2.62 -1.2 ± 0.11 1.7 ± 0.17 1.3 ± 0.03
Nog2 was chosen for further investigation because of the 2.57-fold change at 24
hours (I vs. S) on the microarray. This gene is of further interest because it codes for a
protein that inhibits the bone morphogenic protein during development to result in the
formation of neural tissue. At 3 hours there was no significant microarray data, however
22
a 2.0-fold change (I vs. S) was revealed by qRT-PCR (Table 19). When compared to the
control model we see that both injured and sham-operated fish were up regulated,
however, the injured shows 200% more expression of this gene than the sham-operated
fish. At 24 hours the microarray data revealed a 2.57-fold difference (I vs. S). qRT-PCR
showed a similar 1.6-fold difference (I vs. S), which was due to an increase in the
expression of nog2 in the injured model (1.6-fold I vs. C and -1.1-fold S vs. C). At 168
hours, the microarray showed non-significant results and this was confirmed with qRT-
PCR (1.1-fold I vs. S). There was no differential expression between injured and sham-
operated, however, both sham-operated and injured models showed an increased
expression of 1.4 and 1.5-fold respectively.
Table 19. Nog2 qRT-PCR results. The results of qRT-PCR reflect a similar fold change at 24 hours. The numbers represented for qRT-PCR reflect the fold change calculated with the ∆∆CT method employing GAPDH as the housekeeping gene. The standard error is also reported (± standard error). Comparisons between injured and sham-operated fish are designated (I vs. S), comparisons between sham-operated and control fish (S vs. C), and comparisons between injured and control fish (I vs. C). Ratios with values <1 were converted to fold change by inversion and indicated by a negative sign preceding the number.
Time Point Microarray
(I vs. S) qRT-PCR
(I vs. S) qRT-PCR (S vs. C)
qRT-PCR (I vs. C)
3 hours -- 2.0 ± 0.06 1.9 ± 0.07 3.8 ± 0.04
24 hours 2.57 1.6 ± 0.17 -1.1 ± 0.05 1.6 ± 0.23
168 hours -- 1.1 ± 0.09 1.4 ± 0.15 1.5 ± 0.11
Tubb5 was selected for further investigation based on its differential expression
on the microarray at 168 hours (3.5-fold I vs. S) and its previous implications in nerve
regeneration (Cameron et al., 2005 and Veldman et al., 2007). The up-regulation of
tubulin mRNA’s is often reported in nerve regeneration studies. At 3 hours qRT-PCR
23
showed a 3.1-fold difference between injured and sham models (Table 20). When
compared to control, both the sham-operated and injured fish showed a significant
increase in expression (6.4-fold and 9.5-fold, respectively). At 24 hours qRT-PCR
showed an unexpected 2.1-fold decrease (I vs. S). When compared to control, it reveals
that this is a true decrease in tubb5 expression in the injured fish (-1.6-fold I vs. C). At
168 hours the microarray data indicated a fold difference of 3.5 (I vs. S). The results of
qRT-PCR revealed a much more exaggerated 28.5-fold difference (I vs. S). When
compared to the control, we see this is due to a large increase in expression of the injured
as compared to control. The 2.3-fold decrease in sham-operated fish is unexpected,
however this marked up regulation in the injured model suggests that this gene may play
a role in optic nerve regeneration.
Table 20. Tubb5 qRT-PCR results. The results of qRT-PCR reflect an exaggerated fold change at 168 hours as compared to the microarray. The numbers represented for qRT-PCR reflect the fold change calculated with the ∆∆CT method employing GAPDH as the housekeeping gene. The standard error is also reported (± standard error). Comparisons between injured and sham-operated fish are designated (I vs. S), comparisons between sham-operated and control fish (S vs. C), and comparisons between injured and control fish (I vs. C). Ratios with values <1 were converted to fold change by inversion and indicated by a negative sign preceding the number.
Time Point Microarray
(I vs. S) qRT-PCR
(I vs. S) qRT-PCR (S vs. C)
qRT-PCR (I vs. C)
3 hours -- 3.1 ± 0.20 6.4 ± 0.52 19.5 ± 1.40
24 hours -- -2.1 ± 0.79 1.4 ± 0.09 -1.6 ± 0.07
168 hours 3.50 28.5 ± 2.1 -2.3 ± 0.4 61.4 ± 6.68
Krueppel-like factor 7a (KLF7A) was chosen for qRT-PCR based on research out
of the Goldman lab at University of Michigan in which they showed that KLF6a and
KLF7a were required for axonal sprouting in retinal explants and for optic nerve
24
regeneration by use of morpholino knockdowns (Veldman et al., 2007). This gene was
not represented on the microarray however I wanted to evaluate how their results
compared when using a sham model. In general, KLF7a showed an increased expression
(I vs. S) at all time points with the largest fold change of 4.0 at 24 hours (Table 21). At
all time points the increase in expression is due to an increase in the injured model as
confirmed by (I vs. C). Veldman et al. conducted a time course study spanning 0-24 days
in which they reported 7 days to be the peak expression of KLF7a in the injured fish as
compared to the control (2007). The results presented here indicate a peak at 24 hours,
and therefore, do not concur.
Table 21. KLF7a qRT-PCR results. The results showed an increased expression (I vs. S) at all time points with the largest fold change of 4.0 at 24 hours. At all time points the increase in expression is due to an increase in the injured model as confirmed by (I vs. C). The numbers represented for qRT-PCR reflect the fold change calculated with the ∆∆CT method employing GAPDH as the housekeeping gene. The standard error is also reported (± standard error). Comparisons between injured and sham-operated fish are designated (I vs. S), comparisons between sham-operated and control fish (S vs. C), and comparisons between injured and control fish (I vs. C). Ratios with values <1 were converted to fold change by inversion and indicated by a negative sign preceding the number.
Time Point Microarray
(I vs. S) qRT-PCR
(I vs. S) qRT-PCR (S vs. C)
qRT-PCR (I vs. C)
3 hours NA 1.7 ± 0.16 -1.1 ± 0.04 1.5 ± 0.10
24 hours NA 4.0 ± 0.40 -1.2 ± 0.02 3.5 ± 0.42
168 hours NA 1.7 ± 0.14 1.1 ± 0.09 1.8 ± 0.06
25
CHAPTER IV
DISCUSSION
The experimental design of this project aimed to reveal changes in gene
expression during optic nerve regeneration that may provide a better understanding of the
mechanisms required for successful regeneration of damaged neurons in the CNS. There
is a general increasing trend in the number of genes differentially expressed throughout
the time course, which suggests that earlier genes may be initiating signaling pathways
leading to the response of additional genes later. Genes expressed within the first 24
hours included genes encoding transcription factors, genes involved in chromatin
remodeling and genes implicated in developmental pathways. Genes expressed at the
later time point (168 hours) corresponded to transcription factors, tubulins, ribosomal
subunits, and genes involved in cell metabolism. The complex changes in gene
expression observed support my hypothesis that observing differences in gene expression
between optic nerve injured and sham-operated fish will reveal genes specifically
involved in regeneration.
In several cases, the qRT-PCR did not agree with the microarray data. This is
most evident with YY1 in which the microarray indicated a differential expression of
26
1.85-fold at 24 hours and 2.62-fold at 168 hours and qRT-PCR resulted in a differential
expression of -1.4-fold and -1.2-fold (I vs. S) respectively. These data may call in to
question the quality of the microarray. Reports from the Core Genomics Facility at
Michigan State University indicated the microarray chips had a high background
fluorescence atypical of microarrays commonly used, which may suggest that the arrays
provided by Ocimum Biosolutions were not of the highest quality. However, the
discrepancies between qRT-PCR and the microarray analysis can be explained by the
sensitivities of the two methods. qRT-PCR is a more direct and sensitive method than
microarray analysis and can be considered more reliable.
By comparing the experimental models to a control fish, it was determined that
lunatic fringe and YY1, showed a large increase in expression in both models. This
suggests lunatic fringe and YY1 may show increased expression as a result of stress or
inflammation instead of as a result of neural repair, and it is unlikely that they contribute
to the regeneration of the optic nerve. This is not surprising for YY1 considering that it
codes for a ubiquitously expressed transcription factor (Kurisaki et al., 2003).
There are data for nog2, tubb5, and ATF3 that may suggest a role in nerve
regeneration. Nog2 showed a greater than 1.5-fold increase in expression (I vs. S) for
both 3 hours and 24 hours, but not 168 hours suggesting that this may be an early
response gene. At both 3 hours and 24 hours, it was determined that the fold change
reflected an increase in the injured fish. Nog2 is important in the formation of the neural
plate during development, working as an antagonist to bone morphogenic proteins
(BMPs). BMPs have many actions in the nervous system including cell proliferation,
patterning, cell fate determination, and apoptosis (Mabie et al., 1997). The actions of
27
noggin favor the formation of neural tissue by inhibiting BMPs from interacting with
their receptors (Trindade et al., 1999). The increase in expression of nog2 in the injured
fish may be required to stimulate retinal ganglion cell formation from the retinal stem cell
population. This proposal is substantiated by a study conducted by Setouchi et al.
showed that noggin induced neural precursor cells to differentiate into neurons and
oligodendrocytes (Setoguchi et al., 2004). Most recently, noggin genes have been shown
to promote significant regrowth in corticalspinal tract of mammals when injected at the
spinal cord lesion site (Matsuura et al., 2008).
ATF3 showed >2.0-fold difference (I vs. S) at 3 hours and 24 hours, but no
difference at 168 hours. ATF3 mRNA levels increase greatly in cells when exposed to
stress signals (Hai et al., 1999), which may explain the increase in expression at the early
time points and not at 168 hours. ATF3 codes for a bZIP leucine zipper transcription
factors that bind cAMP response elements to regulate cell proliferation and
differentiation (Hai and Hartman, 2001; Hai et al., 1999). It has also been reported to
serve as an anti-apoptotic and growth-promoting factor for neurons in culture (Nakagomi
et al., 2003). Perhaps the increased expression of ATF3 is promoting retinal ganglion
cell survival and growth after optic nerve injury.
Tubb5 showed a very large increase in expression (I vs. S) at both 3 hours and
168 hours by qRT-PCR. Tubulins are the principle subunits of microtubules, which are
essential to the growth and maintenance of axons (Mitchison and Kirschner, 1988). In
addition to tubb5, several alpha-tubulins represented on the array showed increased
expression in the injured model as compared to sham.
28
I believe that the experimental design of comparing the injured fish to the sham-
operated fish on the array had several benefits. Firstly, we successfully limited the
number of genes that were differentially expressed in the immune system response
category of gene ontology. The gene ontology analysis revealed the immune response
category as one of the smallest represented at all time points. We were also able to show
that in some instances, differences do exist between the sham and control fish (lfng, YY1,
nog2, and tubb5), while for some genes, the sham-operated and control fish show very
similar expression (KLF7a and ATF3).
The microarray data support ways to tie together results from other researchers
investigating regeneration in a variety of models. There are some researchers who believe
CNS regeneration in fish and amphibians is possible due to the re-activation of
developmental pathways. During limb regeneration in newts, the patterning and
formation is regulated by many of the same genes that controlled its initial development,
and the re-expression of these genes is essential to successful regeneration (Candinouche
et al., 1999). The results presented here suggest that the pathways that govern
developmental axon growth and repression may also control regenerative axon growth. It
is possible that the initiation and potentiation of axon regeneration is accomplished
through the reactivation of developmental pathways. However, although aspects of the
developmental pathways are the same, previous studies have indicated that some
signaling mechanisms differ between development and regeneration ( M.Z.A.
Candinouche et al., 1999; Goldman and Ding, 2000; Udvadia et al., 2001). I observed
differential expression of genes involved in 3 developmental pathways: BMP pathway,
notch signal pathway and the Wnt signaling pathway. During the discussion of the qRT-
29
PCR results we suggested the possibility that lfng (notch signaling pathway) showed
increased expression due to inflammatory response and was not specific to neural repair;
therefore, I will not discuss the notch-signaling pathway. The role of nog2 in the BMP
signaling pathway was discussed previously, so I will limit the discussion to the Wnt
signaling pathway.
Wnt Pathway, fzd2 and N-Myc
Wnt proteins are secreted signaling molecules that bind to Frizzled family cell
surface receptors to activate signaling pathways that result in gene transcription (ß-
catenin/Wnt pathway), cell polarization (planar polarity pathway), or an increase in
intracellular calcium (Wnt/Ca2+ pathway). Activation of the canonical Wnt pathway
results in transcription of genes such as the Myc family of transcription factors; including
c-myc which codes for proteins that are transcriptional activating factors, well known
stimulators of cell growth and proliferation. N-myc is another gene targeted for
transcription by the Wnt pathways. N-myc expression has been correlated to
undifferentiated cells in the embryonic kidney, skin and brain; and further differentiation
of these cells requires down regulation of N-myc (Mugrauer et al., 1988; Moens et al.,
1992). There were two N-myc related genes, N-myc downstream regulated gene 1 and N-
myc downstream regulated family member 3a, differentially expressed -1.63-fold at 24
hours and -2.23-fold at 7 days respectively. Recently, activation of Wnt signaling
pathways by application of Wnt3a or inhibitors of GSK-38 has been shown to promote
neural regeneration in mammalian retina through proliferation of Muller glia-derived
progenitor cells (Osakada et al., 2007). Perhaps down-regulation of N-myc through Wnt
signaling pathway is required for differentiation of the retinal stem cell population into
30
new retinal ganglion cells. This conjecture is not far-fetched considering the role of Wnt
signaling in the intestinal stem cell population. In the small intestine of mammals, the
cells of the epithelium are constantly being replaced as a result of cell division from a
population of stem cells in the crypts of the villi. Wnt signaling is responsible for
keeping the stem cells proliferative and the differentiating cells quiescent (Clatworthy
and Subramanian, 2001).
Some scientists believe that the ability of retinal ganglion cells to regenerate
results from the resident stem cell population in the retina. Fish and frogs possess a
proliferative region called the ciliary marginal zone (CMZ) that contains multipotent
stem cells and progenitor daughter cells (Hitchcock and Raymond, 2004). After 60 hours
post fertilization, retinal growth, with the exception of rods and cones, occurs at the CMZ
by the addition of new cells to the retina in concentric rings as long as the eye is growing
(Wehman et al., 2005; Hitchcock et al., 2004). Mammals also possess stem cells at the
CMZ, but their regenerative potential remains unknown. Several studies have researched
the regenerative potential of Müller glia-derived progenitor cells, which have the ability
to differentiate into few retinal cell types (Fausett and Goldman, 2006). I observed
differential regulation between injured and sham fish of several genes (Dicer1 and N-
myc) that are recognized as regulators of stem cell populations in the skin and gut (Moens
et al., 1992; Clatworthy and Subramanian, 2001) respectively that possibly play a role in
the differentiation of the CMZ into retinal ganglion cells.
Gene Ontology Analysis
The analysis of ontological categories of the differentially expressed genes sheds
light on the general physiological processes involved in optic nerve regeneration. The
31
incomplete annotation is a limiting factor in this analysis; of the 722 differentially
expressed genes with a change of at least 1.5-fold, 132 genes had an unknown ontology.
In addition, the oligonucleotides array covers only a portion (14,067 unique probes
representing 8,839 genes) of the approximate 25,000 genes in the zebrafish genome.
Despite these limitations, some conclusions can be made from the available data.
In general, we observe an increase in the number of genes recruited over time
(Table 2). Increased protein synthesis (Sikora-VanMeter et al., 1987) and stimulation of
cytoskeletal genes (Bisby and Tetzlaff, 1992) are characteristic of regenerating neurons.
These observations are confirmed in this study. There is an increase in the number of
genes involved in protein metabolism, particularly those involved in translation
machinery; 70 ribosomal protein genes showed increased expression in the injured as
compared to sham on the microarray. Genes related to cytoskeleton (tubulins and
annexin) also showed an enhanced expression in the injured fish as compared to the
sham.
There was a general temporal decline in expression of genes associated with
phototransduction in the injured model as compared to sham, with the exception of rlbp1
which is essentially not differentially expressed at 168 hours (ratio of 1.19).
Interestingly, there was a -2.08-fold differential expression between experimental and
sham in rhodopsin. Rhodopsin is a rod-specific visual pigment in the retina that initiates
signal transduction when excited by light. This change in expression is similar to results
reported in 2007 by Veldman et al. (Veldman et al., 2007), which showed a fold change
of -1.73 at 3 days when comparing isolated RGC’s from optic nerve crush injuries to
control fish. However, Cameron et al. used rhodopsin as the reference gene for qRT-
32
PCR when comparing control and injured (by patch removal) retinas, which may have
skewed their data (2005). This general decline in genes associated with
phototransduction may result from a decrease in the number of photoreceptors as a result
of apoptosis. It would be interesting to correlate these changes in gene expression to
retina morphology and behavioral studies.
Conclusions and Future Directions
Why is there a difference between mammals and fish in the regenerative capacity
of the optic nerve? In the scope of this project, this question will remain largely
unanswered. The data presented here aimed to observe gene expression changes to gain a
better understanding mechanisms involved in nerve repair in the CNS. The mechanisms
governing optic nerve regeneration are very complex and may require the activation of
developmental pathways for successful axon growth and re-enervation. Whether retinal
ganglion cells in zebrafish become apoptotic after axotomy or survive to sprout new
neurites from the existing cell body is yet to be conclusively determined, although there
is one report of a 20% decline in retinal ganglion cells during regeneration of the
zebrafish optic nerve (Zhou and Wang, 2002). It would be interesting to examine the cell
survival in the zebrafish by TUNEL after optic nerve injury to address that question.
Further inquiry into nog2 and ATF3 would be useful in determining the role of these
genes in optic nerve regeneration in zebrafish. Fluorescence in situ hybridization studies
can be performed to identify the localization of expression within the retina.
Additionally, we observed up-regulation in the injured fish of several genes that
play a role in epigenetic modification, specifically CXXC1 and histone deacetylase 1
(Lee and Skalnik, 2005; Yamaguchi et al., 2005). Future studies can focus on the role of
33
epigenetics in nerve regeneration; particularly whether the differences in chromatin or
histone modifications that exist between mammals and zebrafish offer an explanation as
to why zebrafish have the ability to spontaneously access and regulate genes necessary
for nerve regeneration while mammals do not.
34
APPENDIX
Table 3. Microarray Results: Genes Showing Increased Expression 3 Hours. At 3 hours there were 50 genes that showed increased expression in the injured fish as compared to sham-operated fish. The table below lists the top 20 genes.
3 Hours
Accession Description Ratio
(I VS. S)
P-
value
NM_200931.1 zgc:56065 6.65 0.04
NM_200570.1 selenium binding protein 1 3.35 0.05
NM_200333.1 CXXC finger 1 (PHD domain) 3.09 0.04
AY929292.1 Danio rerio phosphatidylinositol 4-kinase III alpha (pi4kIII alpha)
3.02 0.04
NM_001012262.1 crystallin, gamma S2 2.51 0.02
XM_696168.1 hypothetical protein LOC402822 2.27 0.02
BC076530.1 tumor protein p73-like 2.19 0.02
NM_214773.1 acid phosphatase 5, tartrate resistant 2.06 0.05
NM_001007383.1 zgc:101832 2.05 0.00
AL627263.6 clone RP71-1L9 in linkage group 14 Contains part of a novel gene similar to ATP8B1 (ATPase, Class I, type 8B, member 1), part of a novel gene similar to MCF2 (MCF.2 cell line derived transforming sequence) and two CpG islands
2.05 0.00
NM_200319.1 transmembrane protein 57 1.99 0.02
NM_214716.1 heat shock protein 4, like 1.97 0.03
AF028724.1 zgc:91934 1.96 0.03
XM_682300.1 similar to KIAA0523 protein 1.93 0.04
BX248503.7 clone CH211-232M7 in linkage group 10 1.85 0.04
NM_131147.1 homeo box A11b 1.85 0.05
XM_697090.1 hypothetical protein LOC554904 1.80 0.05
U49412.1 frizzled homolog 2 1.80 0.03
NM_212952.1 ribosomal protein L36 1.77 0.01
NM_212588.1 solute carrier family 20 (phosphate transporter), member 1 1.77 0.02
35
Table 4. Microarray Results: Genes Showing Increased Expression 24 Hours. At 24 hours there were 112 genes that showed increased expression in the injured fish as compared to sham-operated fish. The table below lists the top 20 genes.
24 Hours
Accession Description
Ratio
(I VS. S)
P-
Value
NM_200964 activating transcription factor 3 3.60 0.00
NM_001024811 GTP binding protein 1, like 3.09 0.03
NM_130992 noggin 2 2.57 0.05
NM_130971 lunatic fringe homolog 2.31 0.05
NM_001001399 signal sequence receptor, beta 2.23 0.05
AY178796 annexin A2a 2.19 0.00
XM_687162.2 clone CH211-81A5 in linkage group 19 2.16 0.05
BX000434 CH211-2E18 2.12 0.05
NM_201293 S-adenosylhomocysteine hydrolase-like 1 2.11 0.05
NM_001045083.1 clone DKEY-161L11 in linkage group 2 2.06 0.01
XM_001346372.1 nuclear receptor-related 1 2.04 0.04
XM_692347 similar to Bardet-Biedl syndrome 1 2.00 0.03
NM_207060 transmembrane protein 49 1.98 0.02
XM_691572 similar to BRCA1 interacting protein C-terminal helicase 1 1.94 0.02
XM_695077 similar to conserved hypothetical protein 1.93 0.01
NM_205690 retinaldehyde binding protein 1 1.93 0.02
NM_200281 sarcoma amplified sequence 1.93 0.03
BX511080 clone DKEY-237N7 in linkage group 3 1.92 0.01
NM_131105 alpha-tropomyosin 1.92 0.00
AL928556 clone DKEY-63M7 1.91 0.03
36
Table 5. Microarray Results: Genes Showing Increased Expression 168 Hours. At 168 hours there were 217 genes that showed increased expression in the injured fish as compared to sham-operated fish. The table below lists the top 20 genes.
168 Hours
Accession Description
Ratio
(I VS. S)
P-
Value
NM_200937.1 inhibitor of growth family, member 3 3.96 0.02
BX640466.9 clone CH211-138A11 in linkage group 2 3.52 0.02
NM_198818.1 tubulin, beta 5 3.50 0.00
XM_683892.1 zgc:103738 3.19 0.00
NM_213062.1 ubiquitin-activating enzyme E1 (A1S9T and BN75 temperature sensitivity complementing)
3.12 0.03
CR356231.12 clone CH211-232N7, complete sequence 2.78 0.01
NM_001002378.1 zgc:92066 2.72 0.00
NM_212617.1 YY1 transcription factor 2.62 0.00
NM_200093.1 ORM1-like 1 (S. cerevisiae) 2.59 0.02
NM_130921.1 nonspecific cytotoxic cell receptor protein 1 2.56 0.01
XM_694574.1 similar to ATP-binding cassette, sub-family A member 1 2.54 0.00
NM_131098.1 apolipoprotein Eb 2.54 0.01
NM_200751.1 zgc:73213 2.54 0.01
CR848747.8 clone DKEYP-77H1 in linkage group 16 2.51 0.00
NM_212758.1 peptidylprolyl isomerase A (cyclophilin A) 2.49 0.00
AY391434.1 ribosomal protein SA 2.49 0.00
NM_001004679.1 zgc:103619 2.45 0.02
NM_001007105.1 apoptotic chromatin condensation inducer 1a 2.42 0.02
NM_212756.1 zgc:111860 2.38 0.00
AY394971.1 tubulin, alpha 8 like 4 2.34 0.00
BX548026.10 clone CH211-193D9 2.28 0.02
37
Table 6. Microarray Results: Genes Showing Decreased Expression 3 Hours. At 3 hours there were 17 genes that showed decreased expression in the injured fish as compared to sham-operated fish. The table below lists those 17 genes.
3 Hours
Accession Description
Ratio
(I VS. S)
P-
value
NM_200751.1 zgc:73213 0.44 0.02
NM_200751.1 zgc:73213 0.46 0.01
BX323035.8 clone DKEYP-94H10 in linkage group 2 0.46 0.01
NM_131568.1 transient receptor potential cation channel, subfamily C, member 4 associated protein b
0.47 0.02
NM_213506.1 zgc:63491 0.49 0.03
NM_200090.1 WD repeat domain 75 0.52 0.05
NM_200048.1 arginyl-tRNA synthetase 0.56 0.01
NM_001007063.1 membrane associated guanylate kinase, WW and PDZ domain containing 1
0.58 0.01
NM_205695.1 zgc:77282 0.63 0.04
XM_680501.1 similar to plasma membrane calcium ATPase 0.63 0.02
NM_001017721.1 zgc:112171 0.64 0.05
XM_689534.1 similar to mKIAA0306 protein 0.64 0.05
NM_201471.1 aldehyde dehydrogenase 9 family, member A1 like 1 0.64 0.05
NM_182877.1 solute carrier family 34 (sodium phosphate), member 2b 0.64 0.05
BX649516.8 clone DKEY-51D8 in linkage group 14 0.66 0.03
NM_180965.4 claudin g 0.67 0.04
CR847897.15 clone DKEY-90L8 in linkage group 8 0.67 0.05
38
Table 7. Microarray Results: Genes Showing Decreased Expression 24 Hours. At 24 hours there were 86 genes that showed decreased expression in the injured fish as compared to sham-operated fish. The table below lists the top 20 genes.
24 Hours
Accession Description
Ratio
(I VS.
S)
P-
Value
NM_200410 zgc:64089 0.41 0.02
XM_686959 similar to CG9590-PA 0.44 0.05
XM_679775 similar to cerebellin 2 precursor 0.47 0.01
NM_131594 beta-catenin-interacting protein 0.47 0.02
NM_199946 male germ cell-associated kinase 0.48 0.05
NM_200711 calbindin 2, (calretinin) 0.48 0.00
NM_213364 proteasome (prosome, macropain) subunit, beta type, 3 0.48 0.00
NM_212866 zgc:77051 0.49 0.00
NM_212809 phosphorylase, glycogen; brain 0.49 0.01
BX004766 jagged 2 0.50 0.03
NM_213000 chimerin (chimaerin) 1 0.50 0.01
AF273890 immunoglobulin heavy variable 1-1 0.52 0.00
CR788254 clone DKEY-82K12 in linkage group 2 0.53 0.00
NM_200910 succinate dehydrogenase complex, subunit A, flavoprotein (Fp) 0.54 0.02
NM_213442 serine/arginine repetitive matrix 1 0.54 0.01
NM_131641 paired box gene 6b 0.55 0.03
NM_199926 transcriptional adaptor 3 (NGG1 homolog, yeast)-like 0.55 0.02
NM_001013293 zgc:110753 0.56 0.01
NM_001002299 protein tyrosine phosphatase, receptor type, U 0.56 0.02
NM_001007376 zgc:101877 0.56 0.00
39
Table 8. Microarray Results: Genes Showing Decreased Expression 168 Hours. At 168 hours there were 191 genes that showed decreased expression in the injured fish as compared to sham-operated fish. The table below lists the top 20 genes.
168 Hours
Accession Description
Ratio
(I VS. S)
P-
value
AY050506.1 phosphodiesterase 6G, cGMP-specific, rod, gamma 0.30 0.00
NM_194384.1 aldolase c, fructose-bisphosphate 0.36 0.01
NM_131868.2 guanine nucleotide binding protein (G protein), alpha transducing activity polypeptide 1
0.36 0.02
NM_001007160.1 phosphodiesterase 6A, cGMP-specific, rod, alpha 0.36 0.00
NM_212609.1 guanine nucleotide binding protein (G protein), beta polypeptide 1
0.36 0.00
NM_212755.1 wu:fb12g05 0.36 0.00
NM_131838.2 ATPase, Na+/K+ transporting, beta 2b polypeptide 0.37 0.01
BC091819.1 hypothetical protein LOC553339 0.37 0.03
NM_213202.1 guanine nucleotide binding protein (G protein), beta polypeptide 3
0.37 0.02
BX511094.6 clone CH211-137G12 in linkage group 19 0.39 0.00
XM_686878.1 similar to SI:dZ75P05.1 (novel protein similar to human spindle pole body protein (SPC98P, GCP3))
0.39 0.00
NM_213149.1 FK506 binding protein 5 0.39 0.00
BC076174.1 phosducin 2 0.39 0.00
BC076120.1 opsin 1 (cone pigments), long-wave-sensitive, 2 0.39 0.01
NM_200784.1 coiled-coil-helix-coiled-coil-helix domain containing 2 0.40 0.00
NM_200719.1 ADP-ribosylation factor-like 3, like 2 0.40 0.00
NM_212609.1 guanine nucleotide binding protein (G protein), beta polypeptide 1
0.41 0.00
NM_152955.1 dachshund a 0.41 0.00
NM_001030061.1 transient receptor potential cation channel, subfamily M, member 7
0.41 0.03
BC060894.1 opsin 1 (cone pigments), short-wave-sensitive 1 0.42 0.00
40
Table 9. Gene Ontology: Cell Proliferation. Of the genes differentially expressed at least 1.5-fold or more, 16 were representatives of the cell proliferation category. The ratios at each time point are reported as injury/sham.
Cell Proliferation GenBank Accession
3 Hours
24 Hours
168 Hours
activating transcription factor 3 NM_200964 1.04 3.60 2.12
annexin A2a AY178796 0.96 2.19 2.26
baculoviral IAP repeat-containing 5a AY057057 0.95 1.17 2.12
CCAAT/enhancer binding protein (C/EBP), alpha NM_131885 0.92 1.57 1.67
Dicer1, Dcr-1 homolog (Drosophila) AY386319 0.91 1.63 0.76
E2F transcription factor 4 NM_213432 1.51 1.43 0.85
eukaryotic translation initiation factor 3, subunit 3 (gamma) NM_001003763 0.97 1.22 1.48
fibroblast growth factor 3 NM_131291 1.03 0.66 0.93
guanine nucleotide binding protein (G protein), alpha
transducing activity polypeptide 1
NM_131868 0.90 0.69 0.36
guanine nucleotide binding protein (G protein), beta
polypeptide 1
NM_212609 0.87 0.63 0.36
histone deacetylase 9 NM_200816 1.40 1.25 1.85
neurogenic differentiation NM_130978 0.89 1.21 0.63
similar to Ubiquitin carboxyl-terminal hydrolase 13
(Ubiquitin thiolesterase 13) (Ubiquitin-specific processing
protease 13) (Deubiquitinating enzyme 13) (Isopeptidase T-
3) (ISOT-3)
XM_681175 1.12 1.70 0.96
SWI VS. SNF related, matrix associated, actin dependent
regulator of chromatin, subfamily a, member 4
NM_181603 1.01 0.75 0.66
TNF receptor-associated factor 6 NM_199821 0.95 0.79 0.60
tumor protein p63 BC076530 2.19 1.49 1.53
41
Table 10. Gene Ontology: Axon Extension and Guidance. Of the genes differentially expressed at least 1.5-fold or more, 5 were related to axon extension and guidance. The ratios at each time point are reported as injury/sham.
Table 11. Gene Ontology: Embryonic Development. Of the genes differentially expressed at least 1.5-fold or more, 8 were representatives of the embryonic development category. The ratios at each time point are reported as injury/sham.
Embryonic Development
GenBank
Accession 3 Hours 24 Hours 168 Hours
abl-interactor 1 NM_200738 1.72 0.83 1.27
apolipoprotein Eb NM_131098 1.46 1.07 2.54
beta-catenin-interacting protein NM_131594 0.93 0.47 0.81
frizzled 2 U49412 1.80 0.98 0.88
homeo box A11b NM_131147 1.85 1.08 1.06
jun B proto-oncogene NM_213556 1.20 1.28 1.78
neurogenic differentiation NM_130978 0.89 1.21 0.63
noggin 2 NM_130992 0.93 2.57 0.79
tumor protein p63 BC076530 2.19 1.49 1.53
Axon Extension and Guidance
GenBank
Accession
3
Hours 24 Hours 168 Hours
cadherin 2, neuronal NM_131081 1.17 1.33 1.48
Ephrin a4b (epha4b) NM_153658 0.87 0.65 0.71
fasciculation and elongation protein zeta 1
(zygin I) NM_213396 0.85 0.62 1.01
neural adhesion molecule L1.2 NM_131361 1.14 1.25 1.81
tubulin, alpha 8 like 3 NM_001003558 1.26 1.70 2.12
tubulin, beta 5 NM_198818 1.15 1.01 3.50
tubulin, alpha 8 like 4 NM_200185 1.00 1.50 2.05
42
Table 12. Gene Ontology: Immune System Process. Of the genes differentially expressed at least 1.5-fold or more, 5 were associated with immune system process. The ratios at each time point are reported as injury/sham.
Immune System Process Accession
3 Hours 24
Hours
168 Hours
CCAAT/enhancer binding protein (C/EBP), alpha NM_131885 0.92 1.57 1.67
CXXC finger 1 (PHD domain) NM_200333 3.09 1.03 1.13
Invariant chain-like protein 1 NM_131590 1.19 1.63 1.05
LIM domain only 2 NM_131111 0.78 1.02 0.67
Zgc:91843 NM_001003997 0.91 0.66 0.87
Table 13. Gene Ontology: Neuron Differentiation. Of the genes differentially expressed at least 1.5-fold or more, 9 were representatives of the neuron differentiation category. The ratios at each time point are reported as injury/sham.
Neuron Differentiation
GenBank
Accession
3 Hours 24
Hours
168 Hours
CCAAT/enhancer binding protein (C/EBP),
alpha NM_131885 0.92 1.57 1.67
epha4b NM_153658 0.87 0.65 0.71
neural adhesion molecule L1.2 NM_131361 1.14 1.25 1.81
neurogenic differentiation NM_130978 0.89 1.21 0.63
roundabout homolog 3 AF304131 1.00 0.86 0.65
similar to Bardet-Biedl syndrome 1 XM_692347 0.99 2.00 1.13
similar to Krueppel-like factor 15 XM_688679 1.36 0.62 1.36
similar to plexin C1 XM_685667 0.86 0.62 0.70
similar to SLIT and NTRK-like family, member
4 XM_681309 0.83 0.76 0.42
43
Table 14. Gene Ontology: Phototransduction. Of the genes differentially expressed at least 1.5-fold or more, 16 were associated with photoreception. All but retinaldehyde binding protein are significantly down regulated at 168 hours post injury. The ratios at each time point are reported as injury/sham.
Phototransduction
GenBank
Accession
3
Hours
24
Hours
168
Hours
ATP-binding cassette, sub-family E (OABP), member
1
AY391404 0.96 0.82 0.65
crystallin, gamma S2 NM_001012262 2.51 0.92 0.70
guanine nucleotide binding protein (G protein), alpha
transducing activity polypeptide 1
NM_131868 0.90 0.69 0.36
guanine nucleotide binding protein (G protein), beta
polypeptide 1
NM_212609 0.87 0.63 0.36
guanylate cyclase activator 1B NM_131871 0.96 1.10 0.56
opsin 1 (cone pigments), long-wave-sensitive, 2 BC076120 1.00 0.94 0.39
opsin 1 (cone pigments), short-wave-sensitive 1 BC060894 1.00 0.65 0.42
opsin 1 (cone pigments), short-wave-sensitive 2 NM_131192 1.07 0.97 0.50
phosducin 2 BC076174 0.83 0.64 0.39
phosphodiesterase 6A, cGMP-specific, rod, alpha NM_001007160 0.96 0.63 0.36
phosphodiesterase 6G, cGMP-specific, rod, gamma AY050506 0.95 0.67 0.30
retinal degradation slow 4 NM_131567 0.84 0.88 0.46
retinal homeobox gene 2 NM_131226 1.04 0.86 0.56
retinaldehyde binding protein 1 NM_205690 0.68 1.93 1.19
retinol binding protein 4, like NM_199965 0.99 1.02 0.63
Rhodopsin BC063938 0.88 0.89 0.48
44
REFERENCES Ashburner, M, CA Ball, JA Blake, D Botstein, H Butler, JM Cherry, AP Davis, K
Dolinski, SS Dwight, JT Eppig, MA Harris, DP Hill, L Issel-Tarver, A Kasarskis, S Lewis, JC Matese, JE Richardson, M Ringwald, GM Rubin, and G Sherlock. 2000. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet 25(1): 25-29.
Attardi, DG, and RW Sperry. 1963. Preferential selection of central pathways by
regenerating optic fibers. Exp Neurol 7: 46-64. Beauchemin, M, N Noiseux, M Tremblay, and P Savard. 1994. Expression of Hox A11 in
the limb and the regeneration blastema of adult newt. Int J Dev Biol 38(4): 641-649.
Bernhardt, RR, E Tongiorgi, P Anzini, and M Schachner. 1996. Increased expression of
specific recognition molecules by retinal ganglion cells and by optic pathway glia accompanies the successful regeneration of retinal axons in adult zebrafish. J Comp Neurol 376(2): 253-264.
Beisvag, V, FKR Jünge, H Bergum, L Jølsum, S Lydersen, C Günther, H Ramampiaro,
M Langaas, AK Sandvik and A Lægreid. 2006. GeneTools – application for functional annotation and statistical hypothesis testing. BMC Bioinformatics, 7: 470.
Bisby, MA, and W Tetzlaff. 1992. Changes in cytoskeletal protein synthesis following
axon injury and during axon regeneration. Mol Neurobiol 6(2-3): 107-123. Bustin, SA 2004. A-Z of Quantitative PCR. International University Line Press, La
Jolla, CA. Cameron, DA, KL Gentile, FA Middleton, and P Yurco. 2005. Gene expression profiles
of intact and regenerating zebrafish retina. Mol Vis 11: 775-791. Cao, Y, JS Shumsky, MA Sabol, RA Kushner, S Strittmatter, FP Hamers, DH Lee, SA
Rabacchi, and M Murray. 2008. Nogo-66 Receptor Antagonist Peptide (NEP1-40) Administration Promotes Functional Recovery and Axonal Growth After Lateral Funiculus Injury in the Adult Rat. Neurorehabil Neural Repair 22(3): 262-278.
45
Cho, KS, L Yang, B Lu, H Feng Ma, X Huang, M Pekny, and DF Chen. 2005. Re-establishing the regenerative potential of central nervous system axons in postnatal mice. J Cell Sci 118(5): 863-872.
Clatworthy, JP, and V Subramanian. 2001. Stem cells and the regulation of proliferation,
differentiation and patterning in the intestinal epithelium: emerging insights from gene expression patterns, transgenic and gene ablation studies. Mech Dev 101 (1-2): 3-9.
Clemmons, DR, and LA Maile. 2005. Interaction between insulin-like growth factor-I
receptor and alphaVbeta3 integrin linked signaling pathways: cellular responses to changes in multiple signaling inputs. Mol Endocrinol 19(1): 1-11.
Fausett, BV, and D Goldman. 2006. A role for alpha1 tubulin-expressing Muller glia in
regeneration of the injured zebrafish retina. J Neurosci 26(23): 6303-6313. Fisher, S, and ME Halpern. 1999. Patterning the zebrafish axial skeleton requires early
chordin function. Nat Genet 23 (4): 442-446. Furthauer, M, B Thisse, and C Thisse. 1999. Three different noggin genes antagonize the
activity of bone morphogenetic proteins in the zebrafish embryo. Dev Biol 214(1): 181-196.
Gaze, RM, SH Chung, and MJ Keating. 1972. Development of the retinotectal projection
in Xenopus. Nat New Biol 236(66): 133-135. Gaze, RM, and MJ Keating. 1972. The visual system and "neuronal specificity". Nature
237(5355): 375-378. Gentleman, RC, VJ Carey, DM Bates, B Bolstad, M Dettling, S Dudoit, B Ellis, L
Gautier, Y Ge, J Gentry, K Hornik, T Hothorn, W Huber, S Iacus, R Irizarry, F Leisch, C Li, M Maechler, AJ Rossini, G Sawitzki, C Smith, G Smyth, L Tierney, JY Yang, and J Zhang. 2004. Bioconductor: open software development for computational biology and bioinformatics. Genome Biol 5(10): R80.
Geraudie, J, and V Borday Birraux. 2003. Posterior hoxa genes expression during
zebrafish bony fin ray development and regeneration suggests their involvement in scleroblast differentiation. Dev Genes Evol 213(4): 182-186.
Goldberg, JL, and BA Barres. 2000. The relationship between neuronal survival and
regeneration. Annu Rev Neurosci 23: 579-612. Goldman, D, and J Ding. 2000. Different regulatory elements are necessary for alpha1
tubulin induction during CNS development and regeneration. Neuroreport 11(17): 3859-3863.
46
Hai, T, and MG Hartman. 2001. The molecular biology and nomenclature of the
activating transcription factor/cAMP responsive element binding family of transcription factors: activating transcription factor proteins and homeostasis. Gene 273(1): 1-11.
Hai, T, CD Wolfgang, DK Marsee, AE Allen, and U Sivaprasad. 1999. ATF3 and stress
responses. Gene Expr 7(4-6): 321-335. Hitchcock, P, M Ochocinska, A Sieh, and D Otteson. 2004. Persistent and injury-induced
neurogenesis in the vertebrate retina. Prog Retin Eye Res 23: 183-194. Hitchcock, PF, and PA Raymond. 2004. The teleost retina as a model for developmental
and regeneration biology. Zebrafish 1(3): 257-271. Hsieh, JC. 2004. Specificity of WNT-receptor interactions. Front Biosci 9, 1333-1338. Hunt, D, K Hossain-Ibrahim, MR Mason, and RS Coffin. 2004. BMC Neurosci 5(9)
Sambrook, J., and DW Russell. 2001. Molecular Cloning. Laboratory Manual. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press.
Korenberg, Michael J., ed. 2007. Microarray Data Analysis. Methods and Applications.
Totowa, NJ: Humana Press Inc. Kurisaki K, A Kurisaki, U Valcourt, AA Terentiev, K Pardali, P Ten Dijke, CH Heldin, J
Ericsson, A Moustakas. 2003. Nuclear factor YY1 inhibits transforming growth factor beta- and bone morphogenetic protein-induced cell differentiation. Mol Cell Biol. 13: 4494-510.
Lai, EC 2004. Notch signaling: control of cell communication and cell fate. Development
131: 965-973. Lee, JH, and DG Skalnik. 2005. CpG-binding protein (CXXC finger protein 1) is a
component of the mammalian Set1 histone H3-Lys4 methyltransferase complex, the analogue of the yeast Set1/COMPASS complex. J Biol Chem 280(50): 41725-41731.
Leve, C, M Gajewski, KB Rohr, and D Tautz. 2001. Homologues of c-hairy1 (her9) and
lunatic fringe in zebrafish are expressed in the developing central nervous system, but not in the presomitic mesoderm. Dev Genes Evol 211: 493-500.
Liu, Q, and RL Londraville. 2003. Using the adult zebrafish visual system to study
cadherin-2 expression during central nervous system regeneration. Methods Cell Sci 25(1-2): 71-78.
47
Livak, KJ, and TD Schmittgen. 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods 25(4): 402-408.
Mabie, PC, MF Mehler, R Marmur, A Papavasiliou, Q Song, and JA Kessler. 1997. Bone
morphogenetic proteins induce astroglial differentiation of oligodendroglial-astroglial progenitor cells. J Neurosci 17(11): 4112-4120.
Matsuura, I, J Taniguchi, K Hata, N Saeki, and T Yamashita. 2008. BMP inhibition
enhances axonal growth and functional recovery after spinal cord injury. J Neurochem 105(4): 1471-1479.
Mitchison, T, and M Kirschner. 1988. Cytoskeletal dynamics and nerve growth. Neuron.
1(9): 761-72. Moens, CB, AB Auerbach, RA Cohlon, AL Joyner, and J Rossant. 1992. A targeted
mutation reveals a role for N-myc in branching morphogenesis in the embryonic mouse lung. Gene Dev 6(5): 691-704.
Mugrauer, G, FW Alt, and P Ekblom. 1988. N-myc Proto-Oncogene Expression During
Organogenesis in the Developing Mouse as Revealed by In Situ Hybridization. J Cell Biol 107: 1325-1335.
Nakagomi S, Y Suzuki, K Namikawa, S Kiryu-Seo, and H Kiyama. 2003. Expression of
the Activating Transcription Factor 3 Prevents c-Jun N-Terminal Kinase-Induced Neuronal Death by Promoting Heat Shock Protein 27 Expression and Akt Activation. J Neurosci. 23(12): 5187-5196.
Osakada, F, S Ooto, T Akagi, M Mandai, A Akaike, and M Takahashi. 2007. Wnt
Signaling Promotes Regeneration in the Retina of Adult Mammals. J Neurosci 27(15): 4210-4219.
Rentzsch, F, J Bakkers, C Kramer, and M Hammerschmidt. 2004. Fgf signaling induces
posterior neuroectoderm independently of Bmp signaling inhibition. Dev Dyn 231(4): 750-757.
Ries, A, JL Goldberg, and B Grimpe. 2007. A novel biological function for CD44 in axon
growth of retinal ganglion cells identified by a bioinformatics approach. J Neurochem 103(4): 1491-1505.
Setoguchi, T, K Nakashima, T Takizawa, M Yanagisawa, W Ochiai, M Okabe, K Yone,
S Komiya, and T Taga. 2004. Treatment of spinal cord injury by transplantation of fetal neural precursor cells engineered to express BMP inhibitor. Exp Neurol 189(1): 33-44.
48
Sikora-VanMeter, KC, RC Wierwille, EJ Willetts, and WG VanMeter. 1987. Morphological evidence for increased protein synthesis in CNS neurons after soman exposure. Fundam Appl Toxicol 8(1): 23-38.
Testoni, B, S Borrelli, E Tenedini, D Alotto, C Castagnoli, S Piccolo, E Tagliafico, S
Ferrari, MA Vigano, and R Mantovani. 2006. Identification of new p63 targets in human keratinocytes. Cell Cycle 5(23): 2805-2811.
Trindade, M, M Tada, and JC Smith. 1999. DNA-binding specificity and embryological
function of Xom (Xvent-2). Dev Biol 216(2): 442-456. Udvadia, AJ, RW Koster, and JH Skene. 2001. GAP-43 promoter elements in transgenic
zebrafish reveal a difference in signals for axon growth during CNS development and regeneration. Development 128(7): 1175-1182.
Veldman, MB, MA Bemben, RC Thompson, and D Goldman. 2007. Gene expression
analysis of zebrafish retinal ganglion cells during optic nerve regeneration identifies KLF6a and KLF7a as important regulators of axon regeneration. Dev Biol 312(2): 596-612.
Wehman, AM, W Staub, JR Meyers, PA Raymond, and H Baier. 2005. Genetic
dissection of the zebrafish retinal stem-cell compartment. Dev Biol 281(1): 53-65. Wienholds, E, MJ Koudijs, FJ van Eeden, E Cuppen, and RH Plasterk. 2003. The
microRNA-producing enzyme Dicer1 is essential for zebrafish development. Nat Genet 35(3): 217-218.
Yamaguchi, M, N Tonou-Fujimori, A Komori, R Maeda, Y Nojima, H Li, H Okamoto,
and I Masai. 2005. Histone deacetylase 1 regulates retinal neurogenesis in zebrafish by suppressing Wnt and Notch signaling pathways. Development 132(13): 3027-3043.
Yi, R, MN Poy, M Stoffel, and E Fuchs. 2008. A skin microRNA promotes
differentiation by repressing 'stemness'. Nature 452(7184): 225-229. Zhou, LX, and ZR Wang. 2002. Changes in number and distribution of retinal ganglion
cells after optic nerve crush in zebrafish. Shi Yan Sheng Wu Xue Bao 35(2): 159-162.
VITA
Katherine Elizabeth Saul (Katie) was born to Mary and Gary Saul in 1981. She
was raised in Austin, Texas graduating from James Bowie High School. Katie eagerly
entered Texas State University as a music major in 1999 with the dream of playing cello
alongside Yo-Yo Ma. Her focus shifted to biology in 2002 when her brother was
partially paralyzed from a spinal cord injury. She graduated from Texas State in 2006
with a B.S. in Biology. Katie’s thesis research has focused on nerve regeneration with
the hope of contributing towards the understanding of mechanisms behind reestablishing
functional connections after injury. She is an active member of the community,
volunteering her time as a trainer and equipment manager for a local wheelchair rugby
team.
Permanent Address: 4013 Eskew Drive
Austin, TX 78749
This thesis was typed by Katherine E Saul.